add -j/-p/-w options.

1 files changed, 25 insertions, 9 deletions

diff --git a/uxd.c b/uxd.c
index e839330..b30b22b 100644
--- a/uxd.c
+++ b/uxd.c
@@ -150,6 +150,9 @@ char *dump_data_arg = NULL; /* -d */
 long dump_data_idx  = 0;    /* -d */
 int term_utf8 = 0;    /* -t, -T */
 int restore_term = 0; /* -T only */
+int java_mode = 0;    /* -j */
+int wtf8_mode = 0;    /* -w */
+int permissive = 0;   /* -l */
 
 /* stats for -i option */
 long byte_count = 0;
@@ -265,8 +268,14 @@ void parse_args(int argc, char **argv) {
 			version();
 	}
 
-	while((opt = my_getopt(argc, argv, "tTd:1ic:nbl:rmo:S:s:uhv")) != -1) {
+	while((opt = my_getopt(argc, argv, "jwptTd:1ic:nbl:rmo:S:s:uhv")) != -1) {
 		switch(opt) {
+			case 'j':
+				java_mode = 1; break;
+			case 'w':
+				wtf8_mode = 1; break;
+			case 'p':
+				permissive = 1; break;
 			case 't':
 				term_utf8 = restore_term = 1; break;
 			case 'T':
@@ -551,6 +560,10 @@ int is_overlong(int cont_count, unsigned char *b) {
 	if(!cont_count)
 		return 0;
 
+	/* java mode (MUTF-8) allows exactly one overlong: */
+	if(java_mode && cont_count == 1 && b[0] == 0xc0 && b[1] == 0x80)
+		return 0;
+
 	/* 2 byte seqs, if the first byte is 0xc0 or 0xc1, it's overlong. */
 	if(cont_count == 1 && b[0] <= 0xc1)
 		return 1;
@@ -673,15 +686,18 @@ int dump_utf8_char(void) {
 		}
 	}
 
-	/* don't check bad sequences for out-of-range or surrogate */
-	if(!bad) {
-		if(is_out_of_range(cont_count, bytes) || is_surrogate(cont_count, bytes))
-			bad = 1;
-	}
-
+	if(!permissive) {
+		/* don't check bad sequences for out-of-range or surrogate */
+		if(!bad) {
+			if(is_out_of_range(cont_count, bytes))
+				bad = 1;
+			else if((!wtf8_mode) && is_surrogate(cont_count, bytes))
+				bad = 1;
+		}
 
-	if(is_overlong(cont_count, bytes))
-		overlong = 1;
+		if(is_overlong(cont_count, bytes))
+			overlong = 1;
+	}
 
 	if(bad || overlong) {
 		bad_count++;